HyperDST: Dynamic System-Level Taint Tracking based on Lightweight Virtualization

• Main Authors: Hsieh, Chih-Hao, 謝智皓
• Other Authors: Wu, Yu-Sung
• Format: Others
• Language: en_US
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/wy2hpm

碩士 === 國立交通大學 === 資訊科學與工程研究所 === 108 === Tracking information flow through taint analysis has been a famous research topic for many years. It is helpful for analyzing program behavior, detecting malicious attack, discovering data leakage event and so on. Some approaches transform the program behavior based on the compiler to extract the data flow during runtime. They have higher accuracy but create huge overhead. We proposed HyperDST, which can monitor whole system memory based on lightweight virtualization and perform dynamic taint analysis to trace the information flow of sensitive data. HyperDST collects memory access event and combines all observation results to construct a complete data flow with a certain level of accuracy. Moreover, the overhead on the target application is negligible.